Author Topic: When computers hate you. What computers have refused to work for you?  (Read 4188 times)

0 Members and 1 Guest are viewing this topic.

Offline AmperaTopic starter

  • Super Contributor
  • ***
  • Posts: 2578
  • Country: us
    • Ampera's Forums
So, normally, I'm fairly lucky with computing. I can buy parts, and for the most part, they work. Until my most recent machine, the only parts that have ever really refused to work were a Socket 8 motherboard, and an AWE64 I bought didn't have all the chips needed for anything but PCM audio.

This all changed with my most recent computer, which seems to have had it out for me the day I laid it down.

Now, I intend for this to be a forum about computers in your past (I guess if you need to fit something in that's a bit newer like I did, that's fine too) that have absolutely refused to work, seemingly because they have attained a state of free thought, in which they have decided to hate your guts.

So, without further blathering, let me blather on about what happened.

This machine was intended to fit into the farther end of my line of IBM Compatibles, however inappropriate the term is for something this new. It was to be based around the nForce4 SLI chipset, and have two graphics cards of some sort. I wanted some sort of SATA RAID array, and through looking at all decent options, I had decided that the 939 platform would have fit all my needs the best, as well as a pair of 7800 GTX 512MBs.

Problem 1: First motherboard was completely DOA. Almost all caps were shit, and it appeared to have shit itself through the board. I got my money back here, and could keep the board and the memory it came with. This was a Gigabyte board.

Problem 2: Second motherboard booted, but refused to do anything, but boot into DOS, and reflash BIOS, which didn't end up doing anything else. It would not enter into setup or do anything else, regardless of what I tried. Sent back, money refunded. This was an ASUS board.

Problem 3: Third motherboard. Now, this one isn't really DOA. It was an EPoX 2themax EP-9NPA+SLI, and it actually worked flawlessly for a little while. Little by little, however, it started to gain stability issue. I also had a problem with not getting the SLI selector to work (as these boards have a jumper card to select between dual PCIe 8x, or single PCIe 16x), but this was just because I didn't push down hard enough. This board ended up becoming unstable to the point where I couldn't install windows, regardless of the memory configuration I used. I don't really want to give up on this board, but I guess I have no other choice.

Problem 4: Fourth motherboard. This was the one I ended up using. It's actually not a problem, and works absolutely perfectly. It's an ASUS M2N32-SLI, and it was one my dad was using in one of his machines, just with some random parts in it. I say problem, but that's just to keep a nice list, this works great, but it was a bit too new for what I wanted, as it used DDR2, and an nForce 590-SLI chipset. Why this is worse for me, you have to be my special blend of stupid and crazy to understand.

Problem 5: Fifth motherboard. This was another ASUS board, this time from a neighbor who was awesome enough to give me some load of parts he didn't need anymore. Sadly, this one turned on, but refused to POST.

So, after five motherboards, I'm going with number 4. Don't worry, I tried different sets of RAM (Around 3 different sets) different CPUs (2 different 939 CPUs) and the power supply is brand new and known good (it's working right now.) This, however, was not the end of my problems.

Problem 6: Graphics cards. I have two GTX 7800 512mb editions in this machine, in SLI. I can install Windows XP and drivers, and it will work fine, until I reboot it a few times, at which point it breaks. Device manager reports that both cards can't start up, code 10, and it boots in VESA mode. I previously remedied this (sort of) by uninstalling the drivers, installing an older version, and then reinstalling a new version on top (the older version breaks SLI, which is why I can't use it). This stopped working, and Windows seemed to just refuse to use the cards, for whatever reason.

Long story short, I have found out that it was an IRQ assignment issues. On PCIe. On a semi-modern chipset. How do I know this? Well, the BIOS, as is quite standard, is configured to pass off the assignment of resources and IRQs of devices to the host OS.  This is normal, and something that I have never had break before. Somehow, Windows XP was not assigning proper IRQs to either of the cards, and could only work using, I believe, VESA modes, which can function without specific PCIe drivers. By setting the BIOS to assign IRQs, as if I were using PnP incompatible OS, I have seemed to completely resolve the issue, at least as it stands.

These were just the major issues. I had trouble with drives not showing up, with the graphics cards refusing to connect, severe fan noise, overheating, and just a general case of not wanting to work. It seems to all be set at the moment, and while I have had trouble with it, it's managed to function some of the time, and within those times, it worked quite well.

Also, before anybody says it, I am almost 100% confident none of these are ESD issues. These were handled in dry, non-conductive, mostly wooden, grounded environments, and despite my refusal to use as wrist/ankle strap, I have never had a part die on me for mysterious reasons before or after this. So if it is ESD, it's a consistent run of it happening where it's never happened before, and never happened since.

So, I want to hear the stories of people here. How have computers, old, or possibly new (I consider this a legacy machine, as it's not in common use, not on store shelves, and not powerful enough to be impressively useful for modern tasks) annoyed you, harassed, you, or even just refused to work, and what did you have to do in order to fix it?

I forget who I am sometimes, but then I remember that it's probably not worth remembering.
EEVBlog IRC Admin - Join us on irc.austnet.org #eevblog
 

Offline bob225

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: gb
I have used gigabyte boards for nearly 20 years I probably had 2-3 fail out of 1000's of boards, asus on the other had I had 6 doa (customer supplied) and 10 or more with various faults - mainly voltage reg problems, I will not touch another asus board again - I had major issues with OCZ vertex ssd's - doa, fail to initialise, rma was a mare and then I had 2 doa warranty replacements

I only build for friends and family now as the margins are so tight with out building in volume

 

Offline AmperaTopic starter

  • Super Contributor
  • ***
  • Posts: 2578
  • Country: us
    • Ampera's Forums
Out of all of them, the Gigabyte board was the one that had the most epic failure. ASUS generally is fairly good to me, though. I do agree in general, however, Gigabyte tends to be the nicest company for boards now.
I forget who I am sometimes, but then I remember that it's probably not worth remembering.
EEVBlog IRC Admin - Join us on irc.austnet.org #eevblog
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3893
  • Country: 00
  • STM32, STM8, AVR, 8051
My notebook. For fucked snakes I just can't install the WinUSB driver for the RTL-SDR dongle with R820T2 tuner. (However the other stick with the piece of shit Elonics E4000 works just fine!)  :horse: :horse: :rant: :box:
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23018
  • Country: gb
Recent cockwomble that sits in my mind is trying to get my TTi TF930 talking to my Mac. It uses a prolific USB/serial adapter inside it with a custom vendor and product id. I spent four hours of my life persuading the canned driver that it was supposed to talk to the damn thing. Grr.

Oh and a Corsair survivor USB stick which only works every third time you plug it in.
 

Offline dexters_lab

  • Supporter
  • ****
  • Posts: 1890
  • Country: gb
the quantel hal system i looked at recently was... to put it nicely was 'jolly annoying'  |O

every time i turned the damd thing on i'd get a different set of errors and never got to the bottom of them even after swapping cards for known working ones, there was *always* one disk in the array that was going bad, then a permanent fault appeared on the image card which provided some really nice image corruption on pretty much every operation to add to the image corruption the disk array put on everything. It really was a bit of a basket case.

i described it to it's owner as a grumpy old bit of hardware that doesn't want to be turned on anymore! I should have called it Marvin!


Back in the day i remember buying a PC motherboard with one of the first VIA chipsets, might have been the Apollo Pro so would have been for the old Pentium II etc, i had no end of issues with it and didn't keep it long. After that i vowed to make sure the chipset is the same make as the cpu.

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23018
  • Country: gb
I’d forgotten about VIA. Many an hour wasted on their crap as well.
 

Offline AmperaTopic starter

  • Super Contributor
  • ***
  • Posts: 2578
  • Country: us
    • Ampera's Forums
My only experience with VIA chipsets is with my A-BIT KR7A-RAID Socket A board. It has a VIA chipset, and I have never had an issue with it. It's been quite good, and quite stable, a testament to one of the last members of adherence to what I would call the classic AT/ATX board style, until it was completely replaced by newer 754/939 designs.

A story for another day is my ATAPI CD drive in my 486, which will randomly decide to not like exactly where it is on the controller, and decide that it needs to be relocated to who df knows where in order to work. It's almost as impossible as a Sony/Creative CD controller.

My Pentium Pro machine also had MAJOR bus crosstalk issues. Action on COM ports would induce banding and distortion on the 2D card that was worse depending on what resolution it was at, and would make any modem completely useless. A BIOS update from the OEM gateway BIOS to the Intel BIOS (VS440FX board) most recent solved that issue, and just brought the system into a nice level of stability.
I forget who I am sometimes, but then I remember that it's probably not worth remembering.
EEVBlog IRC Admin - Join us on irc.austnet.org #eevblog
 

Offline Tadas

  • Contributor
  • Posts: 15
Toshiba 1200XE. Dead HD. And dead floppy because of belt becoming a pile of goo after all those years. I disassembled both laptop and floppy drive dozens of times. Replacement belts do not work and it won't boot from CF card since I cannot figure out if there's a BIOS setup in that thing.

It infuriated me so much that I'm not sure I'll pull it out from the storage ever again. Won't throw it away either.

 

Offline Naguissa

  • Regular Contributor
  • *
  • Posts: 114
  • Country: es
    • Foro de electricidad, electrónica y DIY / HUM en español
Few blown random pieces in 25 years or so....

But I'm incompatible with Apple. I can be productive on Windows, Linux, BSD, Solaris, DOS, etc. But I'm very distracted and pushing HW limits (mostly RAM) on MacOS (maybe also on Amiga, but not tested)

Enviado desde mi Jolla mediante Tapatalk


Offline Whales

  • Super Contributor
  • ***
  • Posts: 1899
  • Country: au
    • Halestrom
Much more modern: an ARM based chromebook.  Most of my problems started out software related, but to fix them I was left high and dry by the firmware/lack of BIOS.

Of all the struggles we might have day to day with BIOSs, there has never been a time in my life where I more strongly wished I had one.  Or anything really, anything other than the signed bootloader on the SoC.  That bootloader did nothing other than load another signed bootloader off an 8-pin SOIC.  None of these were reliable, often forgetting settings, locking during boot or not providing basic features until a few power cycles later.

Eventually I had to stop using the laptop because the soldered-on flash had degraded (read: unreliable, hardlocks) and the firmware had decided to refuse to boot off any other devices (eg SD, USB).  The power and body was all there, but the soul was a single point of failure.
« Last Edit: May 15, 2018, 10:58:38 pm by Whales »
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
We once had a Packard Bell (yeah, I know) 386sx-16 back when that was considered a reasonable home computer. It worked fine as originally configured, but when we tried upgrading hard drive everything went to hell in a handbasket. We tried several drives ranging from 120 to 340MB with varying problems. Some drives would work for a bit then start having errors, the 340MB drive would format just fine, I could copy files to it, read files, change directories, install software all no problem. Any time I tried to actually execute anything stored on the drive though the system would instantly hard lock. Putting the original WD 40MB IDE drive back in worked perfectly but no other drives I tried would, didn't matter if they were master, slave, or the only drive. Sometimes I wish I still had that computer so I could put a logic analyzer on it and figure out just what the hell was going on because it bothers me to this day.
 

Offline Whales

  • Super Contributor
  • ***
  • Posts: 1899
  • Country: au
    • Halestrom
Quote
I could copy files to it, read files, change directories, install software all no problem. Any time I tried to actually execute anything stored on the drive though the system would instantly hard lock

Wow.  I wonder if there was some obscure inheritance of memory mapping attributes, eg no execute.  Getting the fault flags out of that processes would indeed have been interesting.

Offline dexters_lab

  • Supporter
  • ****
  • Posts: 1890
  • Country: gb
that reminds me of another issue i had with one of my old Atari TT030 systems a number of years ago, it was a pretty tricked out system but i needed to switch to using the HDDRIVER hard disk driver utility as i wanted to run a different OS on it

the system was utterly reliable in every respect, but as soon as i ran the HDDRIVER app it just bombed out (those who remember Atari ST systems will remember that!) with a bus error, after countless emails back to the developer and removing every single modification i did to the system it still didn't work, yet would run perfectly ok on my spare TT which was identical with the exception of a slightly different system board revision.

I never did get to the bottom of it and put it down to some bizarre hardware bug/fault.

Offline bob225

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: gb
one name that gives me shivers is...……………….

Pc Chips - the single biggest waste of silicon on the planet


Mac OS is nothing more than a glorified front end for unix, Never liked the proprietary food chain or the built in obsolescence either
 

Offline AmperaTopic starter

  • Super Contributor
  • ***
  • Posts: 2578
  • Country: us
    • Ampera's Forums
Shitting on MacOS can be reserved for another pub discussion, unless it's System 9 or earlier.
I forget who I am sometimes, but then I remember that it's probably not worth remembering.
EEVBlog IRC Admin - Join us on irc.austnet.org #eevblog
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23018
  • Country: gb
Yes. Class MacOS deserved it.
 

Offline NivagSwerdna

  • Super Contributor
  • ***
  • Posts: 2495
  • Country: gb
My recent experience in https://www.eevblog.com/forum/microcontrollers/quartus-ii-web-edition-13-1-on-windows-10/msg1404396/#msg1404396

Download a large ,ISO file from the Interweb onto my PC

Running Checksum on the file produces *different* results each time checksum is calculated and file does not extract correctly.

Traced to issue with CMK16GX4M2A2400C14 memory on ASUS Z170-K MB.... took me days of tearing my hair out to work that one out.

Try asking... "My file gives different checksums" in a public forum.... people think you are an idiot.   :)



 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23018
  • Country: gb
ECC FTW!  Been there!

No one thinks they need ECC unit something weird happens.
 

Offline AmperaTopic starter

  • Super Contributor
  • ***
  • Posts: 2578
  • Country: us
    • Ampera's Forums
Yes. Class MacOS deserved it.

I find a certain chaotic charm to it. It had next to nothing to it and pissed everybody off, I never seemed to have a problem with it, and it just sort of worked. I only used it to play a few games on an iMac I have, but it was a neat OS.

My recent experience in https://www.eevblog.com/forum/microcontrollers/quartus-ii-web-edition-13-1-on-windows-10/msg1404396/#msg1404396

Download a large ,ISO file from the Interweb onto my PC

Running Checksum on the file produces *different* results each time checksum is calculated and file does not extract correctly.

Traced to issue with CMK16GX4M2A2400C14 memory on ASUS Z170-K MB.... took me days of tearing my hair out to work that one out.

Try asking... "My file gives different checksums" in a public forum.... people think you are an idiot.   :)


I mean, you do seem a bit like an idiot.  :-DD

Joking aside, that's something I've never had happen. All I've ever had happen was this useless load of crap that is a platform abuse the shit out of me.

ECC FTW!  Been there!

No one thinks they need ECC unit something weird happens.

Besides the fact that this was an issue due to failing memory that would have likely failed entirely later on, ECC isn't available on anything but stupid expensive and limited workstation/server boards (at least from Intel), and it still does nothing for the average consumer, whose most critical task is paying a bill online.
I forget who I am sometimes, but then I remember that it's probably not worth remembering.
EEVBlog IRC Admin - Join us on irc.austnet.org #eevblog
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
I was never a Mac guy back in the day but now I have a handful of vintage Macs and I have to say the OS is pretty slick and the hardware is for the most part beautifully designed, this is of course in comparison to PCs of the same era. I could still fire up MS Word on a Mac II with its stock 640x480 256 color display and it wouldn't look that out of place next to a modern version on a new PC. A 286 running MS-DOS on a CGA or EGA monitor looks absolutely primitive in comparison. The classic MacOS is a lot more polished than early versions of Windows, it wasn't until Win95 that PCs started to really catch up.
 

Offline rdl

  • Super Contributor
  • ***
  • Posts: 3665
  • Country: us
ECC works with a surprising number of cheap Intel and AMD processors. I believe many i3 and some of the cheap dual core Pentiums can use it. My FreeNAS machine is an HP Microserver, which I paid less than $250 for new, has some kind of AMD Turion SoC and it supports ECC RAM.
 

Offline AmperaTopic starter

  • Super Contributor
  • ***
  • Posts: 2578
  • Country: us
    • Ampera's Forums
ECC works with a surprising number of cheap Intel and AMD processors. I believe many i3 and some of the cheap dual core Pentiums can use it. My FreeNAS machine is an HP Microserver, which I paid less than $250 for new, has some kind of AMD Turion SoC and it supports ECC RAM.

Yes, but are the worth using?

Can I get i7 performance at i7 prices with ECC support? Will it come with modern features like NVMe, SLI/Crossfire support, USB 3.1/C, Integrated sound, high speed DDR4, unlocked multipliers? I do not think so.

ECC is worthwhile for people who work with critical data, and can't afford random memory issues, and need better long term stability. It's intended for running consumer data through a machine, handling critical company processes, or for ensuring workstations don't go down as often. They might be nice concepts to have, but consumer memory issues are relatively rare, and are not on a large scale like in enterprise situations.

Anyways. If you wish to discuss this further, DM me, or pop on the IRC and debate people there. We should return to the track of when computers become sentient and begin to hate you.
I forget who I am sometimes, but then I remember that it's probably not worth remembering.
EEVBlog IRC Admin - Join us on irc.austnet.org #eevblog
 

Offline Mr. Scram

  • Super Contributor
  • ***
  • Posts: 9810
  • Country: 00
  • Display aficionado
Re: When computers hate you. What computers have refused to work for you?
« Reply #23 on: July 14, 2018, 04:53:06 am »
ECC makes sense. Consumers are now working with huge sets of data and statistics tells us that the likelihood of wayward bits causing havoc is ever increasing. It's why traditional RAID is becoming obsolete too. ECC should also be more secure.

Why would you need an i7, SLI and NVMe in your NAS anyway?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf